Why Friends Shouldn’t Allow friends to mix and match seasonally adjusted and non-seasonally adjusted data in calculating changes

Reader Bruce Hall commented in his defense of calculating 18 months of change using seasonally unadjusted CPI data, and then 6 months of seasonally adjusted data:

…pgl insists on using 13.3% which is what I wrote in my initial comment regarding “price level change” based on non-seasonally adjusted data. After “correcting my thinking” I used seasonally adjusted Fed data which was 12.6% from January 2021 to June 2022. Of course, we all know that seasonally adjusted data endpoints are subject to revision as seasonal factors are recalculated over time. So maybe after several months that number could be 12.4% or 12.8%. But the issue is not the exact percentage change but the general magnitude of that change and how that relates to the subsequent change (June-December). June seems to be the cutoff point that generally doesn’t show up in year/year numbers, but should be extracted.

Well, I’ve pointed out the risks of making 18-month changes to NSA data here; But what about adding nsa and sa data?

Well, here’s an example (from the Scott Walker campaign, in a 2016 post) of adding not just different strings, but different seasonally adjusted (or unadjusted) data.

In particular, the campaign of Wisconsin Governor Walker ran into trouble when they touted job creation numbers obtained by combining seasonally unadjusted job numbers (from the so-called quarterly census of employment and wages) with seasonally adjusted job numbers (from the Enterprise Survey) to get the change cumulative employment. (They did this because the QCEW numbers are several months behind, while the enterprise survey data is more timely.) This is illustrated in Figure 3.

Figure 3: Non-farm private employment in Wisconsin from the Quarterly Employment and Wage Census, seasonally unadjusted (blue), non-farm private employment from the Enterprise Survey, seasonally adjusted (red). Black arrows indicate changes in QCEW and institutional survey numbers; Teal arrows above the scan of the institution. Source: BLS.

Note that one can calculate the changes from December 2010 (just before Walker took office) to March 2012 (the latest QCEW numbers available as of December 12, 2012), and then add them to the change from March 2012 to October 2012 (the latest institution number available as of December 12 2012). That is, add 89.1 to 6.4, and you get 95.5K, which is close to the 100K figure mentioned by Governor Walker’s campaign. You can see why Governor Walker’s campaign officials did this – the correct calculation using the change in the Foundation survey from December 2010 to October 2012 was only 61.1K.

So, Bruce Hall is, if not in “good” company, then in the company of the likes of Scott Walker’s campaign.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *